As a member of the Big Data Infrastructure group at Salesforce, I worked on scaling, optimizing, and adding new functionality to the Apache Phoenix SQL engine. Following is a brief overview of my work:- Built a framework to add new data encoding capabilities and implement customized projection and value comparison HBase filters that considerably improved query and data load performance across the board.- Wrote a metrics collection framework for collecting and reporting Phoenix internal metrics at a per SQL statement and global level. This framework is extensively used within Salesforce to help tune and monitor the health of the Phoenix/HBase deployment.- Implemented parallelism for a set of queries which reduced their run time by 8x. Also surfaced HBase's native timestamp for queries against time series data which boosted performance varying from 3x to 200x.- Improved Phoenix's statistics collection framework to support query federation and storage counting use cases.- Designed and implemented capability within the Salesforce's platform and Apache Phoenix to efficiently page through data stored in Phoenix/HBase.- Hardened several other core functionalities including secondary indexes, filters, schema changes, etc to ensure they are able to handle enterprise level scale and operational complexity.I also worked for the Database and Cache core infrastructure group where I helped scale Salesforce's relational-data access layer. Some of the work I did there includes:- Addressing scaling issues in core frameworks like app level replication and physical deletes- Removing bottlenecks in the connection pooling layer to reduce average URI time by 10%- Writing tools to help monitor database activity and implementing heuristics that proactively detect resource hogging sessions and connection leaks.

- Worked in the Fraud and Risk organization developing applications and platforms to help prevent professional and buyer/seller fraud, and mitigate and detect associated risk. - Was a member of the User Verification Group and worked on developing and enhancing various User Verification flows. - Also worked for the Release Engineering Team developing applications to automate the various internal processes in a release cycle.

My stint at Samsung as a Software Engineer Intern involved working on performance evaluation of Samsung's proprietary Mobile Ad hoc Network Protocol. I developed performance monitoring tools for Samsung i730 phones running Windows CE. I evaluated the protocol in terms of memory usage, route discovery and route maintenance times, mobility scenarios and performance degradation in grey zones.